What is FinTech?

• The verbatim definition: Financial Technology

FinTech

• The buzzword FinTech started to appear around 2015

• Google Search Trend

• FinTech has its own specific connotation regarding:

  • Business model

  • Area of Finance

  • Technologies applied in Finance

FinTech by Business Model

• FinTech companies expands the customer base by:

  • Democratizing technology

    • through use of mobile apps (e.g. Robinhood, Chime)

  • Large cost-reduction with innovative/streamlined business model

    • through automated models based on big data
    • can have large customer base compared to size of the firm

• B2B, B2C oriented -> P2P, Platform

FinTech by Financial Sector

• Payment systems, Brokerage/Exchange services (Platform)

  • Paypal (Venmo), Zelle

  • Robinhood, Coinbase

• Investments and Portfolio management

  • Robo-advising (Portfolio management)

  • High-Frequency, Algorithmic Trading (Mutual/Hedge Funds)

• Credits and Loans (Banking)

  • Crowdfunding, P2P lending (LendingClub)

FinTech by Technology

The financial companies that use innovative technologies

• Big data
  • by volume & velocity: large, high frequency data
  • by variety: image, sound, text data
• Machine Learning
• Blockchains/ Cryptocurrencies
• Internet of Things
• Virtual Reality / Metaverse

TechFin

FinTech in the Globe

FinTech in Tampa & St. Pete

Our strategic partnership program to grow network of innovators in FinTech world:

• https://www.tampabaywave.org/

AI, ML, DL

Artificial Intelligence

Artificial Intelligence (AI) : intelligent machine

Those machines are called intelligent as:

• They have decision-making capabilities

• like human beings (i.e., “smart”)

Chess and AI

Programming and ML

Machine Learning

Machine Learning (ML): A subset of AI

• A technology to allow computer to learn the patterns and rules
• Without specific instructions

Algorithms make the computer to learn

• Some from Statistics, others from Computer science

• Neural network (Deep Learning) one of such algorithm

Programming and ML

Go (Baduk) and AI

Tesla FSD

Previous versions (V1 ~ 11)

• Programmed approach
• Used hundreds of thousands of hard-coded C++
• Rigid, robotic and jerky driving style

Since 2024 (V12 ~)

• Machine Learning (Deep Neural network)
• Used “safe” human driver data for training
• Improved, natural, human-like driving

AI and ML examples

Finance example: Loans

If loan officers applied certain known rules for loan decisions:

• we ask what rules they applied (should be clear-cut)

• this process can be automated by set of instructions (programming)

• AI (non-ML)

Finance example: Loans (Cont’d)

If we do not know what rules are used, but the outcomes:

• Subjective and fuzzy rules, not always the same rules applied

• We could use ML to learn the rules using data

• ML will make automated decisions based on the data patterns

• AI (ML)

Limits of ML

Complexity-Interpretability Tradeoff

Interpretability Issues

Black Box

• Financial data is inherently complex and interdependant

• Hard to track why a specific decision (e.g., loan approval) was made

• Financial regulators and institutions require clear reasoning behind decisions

• Even harder for complex ML algorithms

Ethical / legal Issues

Example: Discrimination in Loans

• Historical biases in datasets (e.g., lending data) can perpetuate discrimination

• Example: Lower loan approvals for certain demographics due to biased data.

• Will dropping demographc data be sufficient? No.

• c.f., Garbage-In-Garbage-Out problem

AI: The 4th Industrial Revolution

How will it change the financial world?

The Future That Arrived Early

Some insights from how AI shaped the world of Go

• One of the first shocked field that was disrupted by AI (DL)
• A game of longest human history (BC 2200)
• A perfect information game, solvable in theory, not in practice

AI Disruptions

Corner opening moves (joseki)

• AI challenged some traditional, long-accepted patterns (found suboptimal)

• No longer played (deprecated)

• Play focus: imaginative to calculative

  • Convergence of player’s style (meta-game)

AI Empowering Human

AI Empowering Human (Cont’d)

Dramatic improvement in average Human Go skills

• Best in 2012 vs 2016 vs 2020: who wins?

• AI discovers, confirms and strengthens human theory

New AI-based josekis are found and adopted

• Players learn from AI’s playstyle

• Commentators use AI for win rate

Human Learning

AI is less effective for teaching the game for beginner-intermediates:

• It can overwhelm novices
• Can’t understand AI’s complex level
• Often fails to deliver foundational insights
• Ineffiective intuitions

AI becomes more important at high skill levels:

• Far easiaer access to sophisticated level knowledge
• For advanced learners approaching mastery

Insights for Finance

In comparison, the field of Finance is

• A game of incomplete / imperfect information.